Thanks for all your input.  

    I believe that we can avoid the issues of XML having to be translated between different code pages (and thus avoid the problems of XML lying).  There is no reason as to why the XML files can't be UTF-8 on z/OS since there is no specific reason as to why the files have to be human readable.  There are plenty of tools available to convert them as we please.

So really, this is a request for support to be built in to allow XML::Sax / XML::Simple to function correctly on z/OS without having to perform any conversions.  As long as the XML can be parsed then that will be great.  There seems to be enough bits of xml already out there in real unicode (from java and WAS, (and closer to home, broker and AMI)) that being able to parse them as they stand would be a good thing.  ie if the perl parser guessed that it was UTF-8 until it found evidence, such as an encoding attribute in the <?xml tag, that it wasn't.  That would be a big help.  With this support in place, we'd have to ftp the testfiles/*.xml from workstation to USS in binary format - but we'd then expect them to work even though they weren't humanly readable on z/OS.

So essentially, the request is just to make the extensions work on z/OS, without Perl trying to perform any conversions.  Is it possible for this support to be fixed in for the release of Perl-5.8.3?  If not, when can we expect to see it?



Many thanks,


Andrew Boyne
Websphere MQ for z/OS Build Team, DW1H17
Hursley Laboratories
Hursley Park
Tel: 01962 819559
Mobile: 077177 96366
Email: [EMAIL PROTECTED]

Reply via email to